Details
-
Improvement
-
Resolution: Done
-
Major
-
7.2.0
-
0
-
KV 2023-4
Description
There’s a best practice limit that Couchstore buckets shouldn’t get above 1.6TB per Data Service node and Magma buckets shouldn’t get above 16TB per node. As we see issues with rebalance and cluster stability above this number. In particular the individual vBuckets get too large with Couchstore. Prevent a bucket getting larger than this.
Attachments
Issue Links
- is cloned by
-
MB-59113 Self-Managed - Hard limit of bucket data size per node
- Open
- links to
For Gerrit Dashboard: MB-57368 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
193774,21 | MB-57368: Enforce maximum data per bucket per node | master | ns_server | Status: MERGED | +2 | +1 |
194203,13 | MB-57368, MB-57366: Allow for set reason for deny data ingress | master | kv_engine | Status: MERGED | +2 | +1 |
195517,2 | MB-57368, MB-57366: Add 'guard-rail' as errmap tag | master | kv_engine | Status: MERGED | +2 | +1 |
196546,13 | MB-57368: Cluster test for max data per node guard rail | master | ns_server | Status: MERGED | +2 | +1 |